How does the biology community currently feel with regards topublishing descriptive and effect size statistics rather thansignificance stats? Almost every journal article I read in the cellbiology field almost always reports things like P values and statstests to report statistical significance, but should effect sizesbe more important to a biologist? Do we even care if something isstatistically significant if the effect size is negligible? Ratherthan crunch for significance, could one get away with showingthings like confidence intervals, eta^2, Cohen's d, and r valuesinstead over P values? P values tell you the odds that if youassume the null hypothesis is true, then the observation you'remaking are only 5%(assuming of course P<0.05). However, this canlead to the logical fallacy as noted by Aristotle--theory Apredicts that changing X will cause Y. An experimenter thusperforms experiments to manipulate X and sees changes in Y,therefore he/she concludes theory A is supported, which is howevercompletely wrong. Theories B, C, D, E....... could all also predictthat X changes Y and may even be better at it. Even if you concludethat your findings "support" theory A, it's still weak because youhaven't ruled out all of the other possibilities.
So in order to avoid statistical significance relative to nullhypothesis that has all sorts of pitfalls, can one just usedescriptive and effect size statistics just as effectively, if notmore so?
How does the biology community currently feel with regards topublishing descriptive and effect size statistics rather thansignificance stats? Almost every journal article I read in the cellbiology field almost always reports things like P values and statstests to report statistical significance, but should effect sizesbe more important to a biologist? Do we even care if something isstatistically significant if the effect size is negligible? Ratherthan crunch for significance, could one get away with showingthings like confidence intervals, eta^2, Cohen's d, and r valuesinstead over P values? P values tell you the odds that if youassume the null hypothesis is true, then the observation you'remaking are only 5%(assuming of course P<0.05). However, this canlead to the logical fallacy as noted by Aristotle--theory Apredicts that changing X will cause Y. An experimenter thusperforms experiments to manipulate X and sees changes in Y,therefore he/she concludes theory A is supported, which is howevercompletely wrong. Theories B, C, D, E....... could all also predictthat X changes Y and may even be better at it. Even if you concludethat your findings "support" theory A, it's still weak because youhaven't ruled out all of the other possibilities.
So in order to avoid statistical significance relative to nullhypothesis that has all sorts of pitfalls, can one just usedescriptive and effect size statistics just as effectively, if notmore so?
